Hello,

 

Since I upgraded, my server no longer listens on its ipv6 address.
I'd already had this problem with a previous update:

This time it was a bug and was corrected with the next minor version. But here the symptoms aren't exactly the same, so I don't know.

 

The network interface has its own ipv6 and it's up and running:

# ip addr show br0
6: br0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
    link/ether c8:7f:54:a9:c2:1b brd ff:ff:ff:ff:ff:ff
    inet 10.0.2.2/24 metric 1 scope global br0
       valid_lft forever preferred_lft forever
    inet6 2a01:xxxx:yyyy::2/64 metric 1 scope global
       valid_lft forever preferred_lft forever
# ping google.fr
PING google.fr (2a00:1450:4007:810::2003) 56 data bytes
64 bytes from par21s03-in-x03.1e100.net (2a00:1450:4007:810::2003): icmp_seq=1 ttl=112 time=13.0 ms
64 bytes from par21s03-in-x03.1e100.net (2a00:1450:4007:810::2003): icmp_seq=2 ttl=112 time=7.46 ms

 

 

This time, unlike the previous case, my ssh config file is populated with my ipv6 address:

# grep Listen /etc/ssh/sshd_config
ListenAddress 10.0.2.2 # br0
ListenAddress inet6 2a01:xxxx:yyyy::2/64 # br0

 

But the ssh server doesn't listen on it:

# ss -tulpn -4 | grep ssh
tcp   LISTEN 0      128          10.0.2.2:22         0.0.0.0:*    users:(("sshd",pid=1943,fd=6))
# ss -tulpn -6 | grep ssh
#

 

When I restart the ssh service I get no error.

 

 

EDIT:

I just noticed that after I did the /etc/rc.d/rc.sshd restart, as I said I got no error and .. it now is listening on the ipv6:

# ss -tulpn -6 | grep ssh
tcp   LISTEN 0      128       [2a01:xxxx:yyyy::2]:22            [::]:*    users:(("sshd",pid=134056,fd=6))

So I don't know what happened, or if I'll have the problem on the next reboot (in which case I now know how to solve it) but it's solved.
Moderators: you can delete my thread if you think it's no longer of interest. Thank you.

Edited May 21, 20251 yr by pbear
Got the probleme again, after ipgrading to 7.1.2

I had the same problem after upgrading to v7.1.2 : The ssh server does not listen on ipv6 by default. I have to do a manual restart of the ssh deamon to solve the problem.

Fortunately, I don't reboot the server every week, but it's still not normal.

Am I the only one with this problem?

Same problem with v7.1.4:

# ss -tulpn | grep ssh
tcp LISTEN 0 128 10.0.2.2:22 0.0.0.0:* users:(("sshd",pid=1979,fd=6))

# /etc/rc.d/rc.sshd restart
rc.sshd: SSH server daemon... Started.

# ss -tulpn | grep ssh
tcp LISTEN 0 128 10.0.2.2:22 0.0.0.0:* users:(("sshd",pid=3493488,fd=7))
tcp LISTEN 0 128 [2a01:xxxx:yyyy::2]:22 [::]:* users:(("sshd",pid=3493488,fd=6))

Why on earth doesn't the ssh server listen on ipv6 as soon as the server starts up?!

@ljm42

As you helped me last time, and had tracked down an issue with fixed ipv6s, I'd like to ask you if you might have an idea about my rather similar problem. Thank you in advance.

Edited July 1, 20251 yr by pbear

26 minutes ago, pbear said:

Why on earth doesn't the ssh server listen on ipv6 as soon as the server starts up?!

One educated guess of me would be that at start time of sshd (VERY early in the boot sequence) there is no V6 interface loaded and ready already.

You may try with adding "/etc/rc.d/rc.sshd restart" to your go file (maybe with a "sleep 30" ahead?) and see if it works.

Here it works without tricks. but then, maybe because I have a static V6 address? I don't know.

root@F:~# ss -tulpn | grep ssh

tcp LISTEN 0 128 192.168.0.4:22 0.0.0.0:* users:(("sshd",pid=2240,fd=8))

tcp LISTEN 0 128 192.168.4.4:22 0.0.0.0:* users:(("sshd",pid=2240,fd=6))

tcp LISTEN 0 128 [2001:470:XXXX:1::4]:22 [::]:* users:(("sshd",pid=2240,fd=7))

Yes thanks, this addition to the go could be a solution. Afterwards, as I said, I don't reboot the server often, I now know the problem and am always forcibly present when I update or upgrade the hardware, so it's not really blocking. But if there's a bug with the interfaces like you suggest, or something else, it would be cool if someone looked into it and solved the problem. 

It looks like ipv6 isn't very popular :D

(I'm thinking of Unifi, which still doesn't support ipv6 in vpn connections in 2025, grrrr).

17 minutes ago, pbear said:

It looks like ipv6 isn't very popular :D

It's more the lazyness of the people. They had only 30 yrs now to get aquainted to it :-)

Biggest problem that I see is the "lack of NAT" (although, its there if one really needs that #$§!).

I guess, people need to learn it the hard way. Here in Germany many providers only offer V6 connections with "DSLITE" V4 (no backchannel, not usable for incoming services). This forces the guys to move on to V6 only.

But they also have "dynamic prefixes", which is totally unecassary and the most popular router here ("the FritzBox") has a very strange (and sick!) implementation of V6 that drives server admins crazy.

All of this pushes the brake for V6...

Same problem here with 7.1.4. Restart of sshd helped.